Version: 2022 Assembly: Human Dec. 2013 (GRCh38/hg38)
Description
The SCHEMA (Schizophrenia Exome
Meta-Analysis) consortium is an international collaboration that aggregated and harmonized
whole-exome sequencing data to study the role of rare coding variants in schizophrenia.
The dataset includes 24,248 cases and 97,322 controls from diverse global cohorts.
SCHEMA identified genes with exome-wide significant rare variant burden in schizophrenia,
which point to the biology of the disorder.
Data Access
Since the data can be downloaded from the SCHEMA website, and does not seem to be under a license,
we assume that we are allowed to redistribute it in VCF format.
The data can be explored on our website interactively with the
Table Browser or the
Data Integrator.
For programmatic access, our REST API can be used; the
track name is schema.
For bulk download, the VCF file can be obtained from
our download server.
Summary statistics and variant-level results are also available from the
SCHEMA Browser.
Methods
The SCHEMA (Schizophrenia Exome Meta-Analysis) consortium aggregated whole-exome sequencing
data from 24,248 schizophrenia cases and 97,322 controls (including non-psychiatric,
non-neurological samples from the gnomAD consortium) across multiple international cohorts.
Exome sequencing was performed using various capture platforms and Illumina sequencing
instruments across cohorts sequenced over approximately a decade. Sequence data were
uniformly reprocessed through the BWA-Picard-GATK best practices pipeline as part of the
gnomAD v2 infrastructure, including alignment to GRCh37/hg19, duplicate marking, base
quality score recalibration, and per-sample variant calling with GATK HaplotypeCaller,
followed by joint genotyping across all samples. A novel exon-by-exon coverage estimation
pipeline was developed to account for differences in capture technology across sequencing
batches, and both site-level and genotype-level quality filters were applied. Protein-truncating
variants (PTVs) were annotated using LOFTEE (Loss-Of-Function Transcript Effect Estimator),
and missense variant deleteriousness was scored using MPC (Missense badness, PolyPhen-2,
and Constraint). Gene-level association testing combined: (1) a case-control rare variant
burden test aggregating ultra-rare PTVs (Class I: PTV and MPC > 3; Class II: missense
MPC 2–3) across 18,321 protein-coding genes; and (2) de novo variant enrichment
from 3,402 schizophrenia proband-parent trios assessed via a Poisson rate test against
gnomAD-derived baseline mutation rates; with the two components combined using a weighted
Z-score meta-analysis. This identified 10 genes at exome-wide significance (P < 2.14
× 10-6) with odds ratios for PTVs ranging from 3 to 50, and 32 genes at
FDR < 5%. Full data are available at
schema.broadinstitute.org
(Singh, Neale, Daly & the SCHEMA Consortium,
Nature 2022).
We downloaded the TSV data from the SCHEMA website
and converted it to VCF format using a custom Python script. The VCF was lifted to hg38 using our hg19ToHg38 chain
file.
We provide documentation that indicates how all source files of the varFreqs track were converted in the makeDoc file of the track.
For some tracks, python scripts were necessary and are also available from GitHub.